Cambridge UK-based start up Optalysys has stated that it is only months away from launching a prototype optical processor with the potential to deliver exascale levels of processing power on a standard-sized desktop computer. The company will demonstrate its prototype, which meets NASA Technology Readiness Level 4, in January of next year.
Alhough it’s only at proof-of-concept stage, the processor is expected to run at over 340 gigaFLOPS – enabling it to analyze large data sets, and produce complex model simulations, in a laboratory environment.
Founder and CEO Dr Nick New explains, “Optalysys’ technology applies the principles of diffractive and Fourier optics to calculate the same processor intensive mathematical functions used in CFD (Computational Fluid Dynamics) and pattern recognition.
Although the technology has the potential to deliver several quintillion floating point operations per second (exaFLOPs), it can be powered from a standard mains supply, with an estimated running cost of just £2,100 per annum.
By comparison, the world’s fastest supercomputer, the $320m Tianhe-2 manufactured by China’s National University of Defence Technology, costs a hefty $21m per year to run.
A similarly powered Optalysys supercomputer would be far cheaper to build too.
Nick continues, “We are currently developing two products, a ‘Big Data’ analysis system and an Optical Solver Supercomputer, both of which are expected to be launched in 2017.
“The analysis unit is designed to work alongside existing supercomputers to provide advanced data analysis capability. Its initial specification will be 1.32 petaFLOPs, increasing to more than 300 petaFLOPs by 2020.
“The Optalysys Optical Solver Supercomputer will initially operate at a speed in excess of 9 petaFLOPs, increasing to 17.1 ExaFLOPs by 2020.”
Putting that into context, current global supercomputing leader Cray is aiming to break the 1 exaFLOP threshold in the same timeframe.
Unlike today’s supercomputers, which necessitate dedicated sites and climate controlled environments, the Optalysys computer will be able to operate in a conventional office setting too.
Optalysys Chairman James Duez says, “In science, engineering and ‘Big Data’ applications — such as aerofoil design, weather forecasting, MRI data analysis and quantum computing — it is becoming increasingly common for traditional computing methods to fall short of delivering the processing power needed.
“The optical processing technology that we are developing will offer step-changing advances in how physical processes are modelled and how the resulting data sets are analyzed.”
James continues, “Our initial target market is CFD, an essential tool for a quarter of a million engineers and scientists around the world.
“CFD can be used to predict the weather, design cars and model airflow, but the speed of processing needed to create models is constrained by current electrical capabilities.
“The speeds that optical processing can achieve will eliminate this problem almost overnight.”
Nick says, “Early conversations with potential customers have been extremely positive and one of the largest weather centers have said they are keen to collaborate with us because the energy cost to produce such high quality forecasts, and deal with the huge data volumes, is unaffordable with current processor technologies.
A short film explaining this revolutionary technology, narrated by Professor Heinz Wolff, is available to view here.
For more information, visit www.optalysys.com.